Does an underactive thyroid explain weight gain?
Partly, and less than most people expect. The weight attributable to an underactive thyroid is roughly 2 to 5 kg, most of it retained salt and water rather than fat[1]. Treatment does not reliably return it: of newly treated patients, about half lost weight over two years, averaging 3.8 kg, and half gained[2]. Get the gland treated, then treat the weight as a separate problem.
On this page
What thyroid hormone actually sets
The thyroid sits at the front of the neck and releases two hormones, thyroxine and triiodothyronine, whose job is to set the pace of almost every tissue in the body. Heart rate, gut transit, body temperature, how fast cells burn fuel at rest — all of it is tuned by how much of that hormone is in circulation. When there is too little of it the condition is called hypothyroidism, and the symptoms follow the same logic: cold, tired, constipated, slow.
Because one of the things thyroid hormone sets is resting energy use, it is reasonable to expect an underactive thyroid to show up on the scale. It does. The question that matters to someone newly diagnosed, or suspicious that they should be, is a quantitative one: how much of the weight is the gland, and how much is everything else? The answer is smaller than most people are told.
How much weight the thyroid can account for
The American Thyroid Association puts the weight attributable to an underactive thyroid at roughly 2 to 5 kg, depending on how severe the deficiency is, and notes that most of it is retained salt and water rather than fat[1]. That second half is the part that gets lost in the telling. Low thyroid hormone slows the kidney's handling of sodium and changes the tissue under the skin so that it holds fluid, which is why an untreated patient's face and hands look puffy, and why the gain often arrives faster than any plausible calorie surplus could explain.
So somebody who has put on twenty kilograms over six years and then tests positive for hypothyroidism has found a real diagnosis and a partial explanation. Treating it is not optional; the condition has consequences well beyond bodyweight. But the arithmetic behind the rest of the gain is not changed by the diagnosis.
Why treatment does not usually hand the weight back
This is where expectation and evidence separate most sharply. Follow-up of patients newly diagnosed with primary hypothyroidism and brought back to normal thyroid levels on levothyroxine found that only about half of them lost any weight at all over the following two years, and that those who did lost an average of 3.8 kg[2]. The other half gained. Restoring the hormone clears the fluid and removes the drag the gland was adding; it does not undo years of habits that formed while the gland was underactive.
The related trap is treating a borderline result. A clinical practice guideline that reviewed 21 randomised trials in more than two thousand people concluded that thyroid hormone for subclinical hypothyroidism — a mildly raised TSH with normal thyroid hormone levels — produced no meaningful improvement in symptoms, fatigue or quality of life, and recommended against it for most adults[4]. A tablet started in the hope of weight loss is the weakest reason there is to start one.
Is the metabolism permanently slower?
Partly, and by less than the word "slow" suggests. Women being treated for hypothyroidism, with a TSH already back in range, measured a resting energy expenditure of 28.59 kcal per kilogram of fat-free mass per day against 29.91 in matched women without the condition — a gap of about 4%[3]. It is a real difference and a statistically solid one. It is also about what a short walk costs. It is not the difference between losing weight and not losing weight.
Two things follow. The first is that a treated thyroid does not need a special diet; it needs the same energy balance as everybody else, worked out the ordinary way. The second is that a shortfall of about 4%[3] is precisely the sort of thing a target built from measured bodyweight absorbs by itself, because lower resting expenditure produces slower loss, and a target recalculated every few kilograms follows that loss down. The guide on how your daily calories are calculated shows where the number comes from, and how fast you can safely lose weight covers what a realistic rate looks like once it is set.
What to do with the diagnosis
Get tested if the symptoms are there: cold you cannot shake, fatigue that sleep does not fix, dry skin, constipation, a hoarse voice, weight that appeared without an explanation. It matters more if you are a woman past forty or have a family history, because that is who this happens to most often. A TSH with a free thyroxine is a cheap blood test and its answer is unambiguous.
Then separate the two problems. The gland belongs to your doctor, and it is treated with a tablet and a follow-up test, not with food. The weight is a food and training problem, and it stays one whether or not the gland is cooperating. If the part you actually struggle with is knowing what you ate this week, that is a tracking and accountability gap rather than an endocrine one, and it is what an online nutrition coach is for: something that sees the week instead of the morning.
The useful way to hold all of this is that an underactive thyroid makes losing weight harder the way a headwind makes cycling harder. It is real, it is measurable, and it does not change the direction you have to pedal.
Sources
- American Thyroid Association. Thyroid and Weight (patient information brochure). ↩
- Lee SY, Braverman LE, Pearce EN. Changes in body weight after treatment of primary hypothyroidism with levothyroxine. Endocrine Practice 2014;20(11):1122–1128. ↩
- Muraca E, Ciardullo S, Oltolini A, et al. Resting energy expenditure in obese women with primary hypothyroidism and appropriate levothyroxine replacement therapy. The Journal of Clinical Endocrinology & Metabolism 2020;105(4):e1741–e1748. ↩
- Bekkering GE, Agoritsas T, Lytvyn L, et al. Thyroid hormones treatment for subclinical hypothyroidism: a clinical practice guideline. BMJ 2019;365:l2006. ↩